Compare Gene Importance Across Datasets Using Random Forest — calculateVarImpOverlap • scDiagnostics

This function identifies and compares the most important genes for differentiating cell types between a query dataset and a reference dataset using Random Forest.

calculateVarImpOverlap(
  reference_data,
  query_data = NULL,
  ref_cell_type_col,
  query_cell_type_col = NULL,
  cell_types = NULL,
  n_tree = 500,
  n_top = 50
)

Arguments

reference_data: A SingleCellExperiment object containing numeric expression matrix for the reference cells.
query_data: A SingleCellExperiment object containing numeric expression matrix for the query cells. If NULL, then the variable importance scores are only computed for the reference data. Default is NULL.
ref_cell_type_col: A character string specifying the column name in the reference dataset containing cell type annotations.
query_cell_type_col: A character string specifying the column name in the query dataset containing cell type annotations.
cell_types: A character vector specifying the cell types to include in the plot. If NULL, all cell types are included.
n_tree: An integer specifying the number of trees to grow in the Random Forest. Default is 500.
n_top: An integer specifying the number of top genes to consider when comparing variable importance scores. Default is 50.

Value

A list containing three elements:

var_imp_ref: A list of data frames containing variable importance scores for each combination of cell types in the reference dataset.
var_imp_query: A list of data frames containing variable importance scores for each combination of cell types in the query dataset.
var_imp_comparison: A named vector indicating the proportion of top genes that overlap between the reference and query datasets for each combination of cell types.

Details

This function uses the Random Forest algorithm to calculate the importance of genes in differentiating between cell types within both a reference dataset and a query dataset. The function then compares the top genes identified in both datasets to determine the overlap in their importance scores.

References

Breiman, L. (2001). "Random forests". *Machine Learning*, 45(1), 5-32. doi:10.1023/A:1010933404324.

Author

Anthony Christidis, anthony-alexander_christidis@hms.harvard.edu

Examples

# Load data
data("reference_data")
data("query_data")

# Compute important variables for all pairwise cell comparisons
rf_output <- calculateVarImpOverlap(reference_data = reference_data, 
                                    query_data = query_data, 
                                    query_cell_type_col = "SingleR_annotation", 
                                    ref_cell_type_col = "expert_annotation", 
                                    n_tree = 500,
                                    n_top = 50)

# Comparison table
rf_output$var_imp_comparison
#>              CD4-CD8     CD4-B_and_plasma          CD4-Myeloid 
#>                 0.70                 0.80                 0.58 
#>     CD8-B_and_plasma          CD8-Myeloid B_and_plasma-Myeloid 
#>                 0.72                 0.54                 0.56